Wither CASA?
Malcolm Slaney
(IBM Almaden Research Center)

What has happened to CASA and will it solve our problems? Like many researchers, I spent years building a system that uses bottom-up cues to segment the acoustic environment. We have exquisitely sensitive tools, such as the correlogram, to model auditory perception. Yet, computational auditory scene analysis remains a distant dream. In many ways, we haven't come that far since Mitch Weintraub's first CASA system back in 1984. I believe this is caused by our desire to re-synthesize the individual sounds, a task which sounds obvious but is unlikely to be done by the human brain. This talk will summarize the work to date and will discuss some new top-down directions that appear promising.

Relevant material:

Malcolm Slaney (1997). A Critique of Pure Audition. In: Computational Auditory Scene Analysis. Dave Rosenthal and Hiroshi Okuno (Eds.).Mahwah, NJ : Lawrence Erlbaum Associates, pp. 16-30. http://rvl4.ecn.purdue.edu/~malcolm/interval/1997-056/PureAudition.pdf
Malcolm Slaney (1995). Pattern playback from 1950 to 1995. Paper presented at the 1995 IEEE Systems, Man and Cybernetics Conference, October 22-25, 1995, Vancouver, Canada. http://rvl4.ecn.purdue.edu/~malcolm/interval/1994-036/IEEE-SMC-95.pdf